however, longer articles might be expected, but we in adjacent but less related disciplines. Computerised found no supporting evidence. This may suggest that searching based on key words, which in some cases editors are trying to include more contributions in the includes a search only within the title, may also available space. Alternatively, narrowing the range of. encourage this trend. material covered or reporting less experimental work Could there, however, be another explanation? The BMJ: first published as 10.1136/bmj.297.6664.1632 on 24 December 1988. Downloaded from in a given article may be responsible, perhaps a correlation between the length of title and the number reflection of "publish or perish." These factors, of authors but not with the other variables raises a together with the rising number of journals, may place question. Substantive discussion among the increasing increasing pressure on the interested reader, who may number of authors per article might have been not find time for adequate consideration, even of expected to lead to longer articles influenced by a abstracts. Could it be, then, that the longer titles are an greater diversity of viewpoints. This is not the case. attempt to draw attention to the.kernel of the material Our evidence questions whether active contribution by in a sort of mini abstract? This might be valuable in some participants is reserved for the title, which alone attracting the more marginal attention of staffworking bears the thumbprint ofall.

The haggis tolerance test in. Scots and Sassenachs

Alan G Fraser, Alan Rees, Stephanie Matthews, Geraint T Williams

Abstract assigned according to the country of birth of both To find out if the Scottish national dish, haggis, parents. contributes to the high incidence of coronary heart Ethical approval was not sought, and informed disease in Scotland the lipaemic effect of a meal of consent could not be obtained as so little is known 200 g of haggis was measured in six Scottish and about the consequences ofeating haggis. There was no 10 Sassenach men. The Scots had higher fasting control group because we deemed it unethical to cholesterol and triglyceride concentrations and a withhold haggis from any subject. lower proportion of high density lipoprotein choles- After the subjects had fasted overnight and terol than the Sassenachs. Four subjects were found abstained from alcohol blood samples were taken for to have hyperlipoproteinaemia, which had been measurement of fasting lipid concentrations. All sub- unrecognised previously. Serum cholesterol con- jects then ate 200 g offreshly boiled haggis (T G Willis, centrations did not change after haggis was eaten Edinburgh) as their breakfast. Their compliance was (mean dose 2-6 gfkg body weight). Serum concentra- ensured by all subjects eating together. Apple or tions of triglycerides increased by 51% at 90 minutes orange juice was permitted to help swallowing, but in the Sassenachs but were unaltered in the Scots. whisky was withheld until the end of the study. Blood There were.no serious adverse effects. samples were taken 30 and 90 minutes after the last This study shows that Scots have higher lipid con- mouthful of haggis. Subjects were given diaries in centrations than Sassenachs but seem to be resistant which to record any-unwanted symptoms. http://www.bmj.com/ to the lipaemic effect ofhaggis. The haggis tolerance The haggis that we used is commercially available test may be useful in Sassenachs. and prepared in bulk (0 5 tonnes a week) according to traditional recipes.2 Its ingredients are lambs' pluck (heart and lung), pigs' liver, beefsuet (perinephric fat), Introduction and pinhead (coarse) oatmeal. These are cooked; Coronary heart disease is now the main cause of minced; seasoned with salt, black pepper, coriander, death in Scotland. One important contributing factor and ginger; and boiled in a sheep's paunch (stomach). is consumption of fat, but, strangely, the influence of A 200 g portion of haggis contains 21-4 g protein, on 23 September 2021 by guest. Protected copyright. the Scottish national dish, haggis, on blood lipid 43-4 g fat, 38-4 g carbohydrate, 5% inedible fibre, and concentrations and heart disease seems to have been 2584 kJ.3 neglected as a subject for serious scientific inquiry. Quality was checked by performing histological We are aware of only two previous medical investi- studies on a generous wedge biopsy specimen excised Departments of gations ofhaggis. Single blind histological studies per- from a randomly selected haggis (fig 1). Electron Cardiology, Medicine, formed by Professor Horst Oertel at McGill University Medical Biochemistry, and in 1929 confirmed that haggis consists of degenerate Pathology, University liver cells and cereal products within a sac lined by Hospital of Wales, Cardiff epithelial and muscle tissue (J R Cameron, personal CF4 4XN communication). A paper delivered to the Nutrition Alan G Fraser, MRCP, senior registrar Society extolled the virtues of haggis, turnips, and Alan Rees, MRCP, lecturer potatoes as a healthy and balanced diet.' An exhaustive Stephanie Matthews, search of publications produced no other studies, and MRCPATH, haggis is not cited in Index Medicus. We therefore senior registrar set up the Cardiffheart, alveoli, gut and grain ingestion Geraint T Williams, study (HAGGIS) to investigate the effects of con- MRCPATH, senior lecturer sumption ofhaggis on blood lipid concentrations. Correspondence to: Dr A G Fraser, Department of Subjects and methods Echocardiography, Thoraxcentrum Ba 302, We studied 16 healthy male volunteers (that is, Dijkzigt Hospital, 3000 DR conscripted medical colleagues), including six expat- Rotterdam, The riate Scots and 10 Sassenachs. Their mean age was 36 FIG 1 -Haggis cold and before gasirotomy, showing gastric blood Netherlands. (95% confidence interval 31 to 41). N-ationality was vessels

1632 BMJ VOLUME 297 24-31 DECEMBER 1988 BMJ: first published as 10.1136/bmj.297.6664.1632 on 24 December 1988. Downloaded from http://www.bmj.com/ on 23 September 2021 by guest. Protected copyright.

FIG 2-Sectiomns of tissues in haggis stained with haematoxylin and eosin (a) lung; (b) liver; (c) cardiac muscle; (d) skeletal muscle; (e) monocotyledenous endosperm with embryo

microscopy of the specimen was not feasible owing to determined by rate nephelometry (Beckman Immuno- its particulate nature, but conventional examination chemistry System). confirmed the presence of pulmonary, cardiac, and To conform with current standard practice' we used hepatic tissue as well as seeds of the graminae family, those statistical tests that produced significant results. presumably oatmeal (fig 2). Unexpectedly the sample Differences in demographic characteristics between also contained striated muscle; although this is not a the groups were ignored. One Sassenach who had not recognised ingredient of haggis, small quantities -of fasted had high serum triglyceride concentrations and diaphragm adjacent to heart, lung, or liver seemed to was excluded from the analysis oftriglyceride and high have become incorporated in the haggises used in this density lipoprotein cholesterol concentrations. study. Nevertheless, this is unlikely to have seriously influenced the results. Serum cholesterol concentration was determined Results by an enzymatic colorimetric method (Boehringer The table gives the results of the lipid assays. The Mannheim). Concentrations of high density lipo- Scots had higher fasting total cholesterol and tri- protein cholesterol and subfractions were measured by glyceride concentrations than the Sassenachs and a the method of Gidez et al.4 Serum concentrations of lower proportion of high density lipoprotein choles- triglycerides were measured by a Technicon RA1000 terol. Differences in other variables, including the analyser with a kit from Technicon Diagnostic. Serum ratio of apolipoprotein AI to apolipoprotein B, which concentrations of apolipoproteins AI and B were correlates with the extent of coronary artery disease,6

BMJ VOLUME 297 24-31 DECEMBER 1988 1633 Mean lipid concentrations in six Scots and nine Sassenachs (except consumption and heart disease over the centuries where stated otherwise). V'alues in parentheses are 95% confidence and studies of Scottish migrants would be valuable. intervals Regular consumption of haggis may produce a dietary p Value tachyphylaxis with reduced chylomicronaemia-for (unpaired t example, by autoregulation of intestinal haggis Scots Sassenachs test) receptors (should they exist) or derepression ofalterna- BMJ: first published as 10.1136/bmj.297.6664.1632 on 24 December 1988. Downloaded from Cholesterol (mrnol/l): tive catabolic pathways for haggis. Sassenachs may be Fasting 6 5 (5-0 to 7 9) 4-8 (4 0 to 5 6)* 0-037 slow metabolisers of haggis. The Scot who showed the At 30 minutes 6-6 (5 0 to 8-2) 4-8 (3 9 to 5 6)* 0 037 At 90 minutes 6-6 (4-9 to 8-4) 4-8 (4O0 to 57)* 0-049 greatest postprandial fall in triglyceride concentration Triglycerides (mmol/I): spent the interval between the 30 and 90 minute Fasting 1l4(I 0to 1-8) 0 7(0-4to0 9) 0-006 At 30 minutes 1 5 (0-9 to 2 0) 0-8 (0-5 to 1-2) 0-033 samples performing cardiac surgery, which suggests At 90 minutes 1l4 (0 7 to 2 1) 1 1 (0-7 to 14) NS that this may have a protective effect, but the practice HDL cholesterol of cardiac surgery could hardly be advocated as a mass (mmolfl) 1 1 (0-9 to 1-3) 1-2 (11 to 13) NS HDL: total preventive measure. cholesterol 0 18 (0l13 to 022) 0-27 (0-24 to 0 30) <0 0001 For Scots the clear implication of this study is that HDL2 HDL3 0-83 (0-47 to 1-18) 0-69 (0-56 to 0-83) NS Apolipoprotein Al: they should continue to eat haggis, especially if it apolipoprotein B 1l5 (0-8 to 2 2) 2-0 (1 7 to 2 2)* NS is baked rather than boiled as this reduces its fat HDL= High density lipoprotein. content (unpublished personal observations). Haggis *Mean value in 10 subjects. is a healthy food-for example, sheep's lung contains only 20% ofthe fat and more protein weight for weight were not significant. Three subjects were found to have than a leg of lamb.8 Haggis has been recognised for hypercholesterolaemia, which had been undetected centuries as the greatest sausage9 and as "real food that previously. They had cholesterol concentrations nourishes both through feeling and fact,"'0 even if it between 6-7 and 9-2 mmol/l, which are above recom- has on occasion to be disguised as mince in order to mended limits.7 One of these subjects and one other deceive and delight the average southerner." Queen had raised triglyceride concentrations. Victoria attested to its beneficial effects when she ate Ingestion of haggis caused no change in cholesterol "the celebrated haggis" at Blair Castle on Wednesday concentrations in either group (paired t test) so the 11 October 1865 and "really liked it very much."'2 difference between the groups was the same at 90 Perhaps the Scots should export it to countries with a minutes. Overall, haggis caused a significant increase high incidence of coronary heart disease and campaign in triglyceride concentrations, which rose from a for the establishment of haggis subsidies and haggis fasting value of 1 1 (95% confidence interval 0 7 to mountains throughout the European Community. Finally, we suggest that the haggis tolerance test 0 4- Sasseriac s 1 4) mmol/l to 1-3 (0-8 to 1-7) mmol/l at 30 minutes (p=0-012) and 1 3 (0 9 to 1-7) mmol/l at 90 minutes as described in this study offers a novel test for 02- (p=0027), representing an average increase of 34%. fat tolerance, although caution may be required in The rise was, however, confined to the Sassenachs (fig patients with heart failure because 200 g of haggis contains 1 540 mg ofsodium.3 The tolerance test i Scots 3). At 90 minutes, the Scots' triglyceride concentra- haggis tions had increased on average by only 6%, compared is more gastronomically pleasing than previous alter- 0-2 ______natives consisting of butter, double cream, and ice 0 60 90 with an average increase in the Sassenachs of 51% 30 cream, and it is safer and more than tests Time minutes) (p<0-0001). Increases in the concentration of tri- physiological FIG 3-Mean (SE) changes in glycerides were not related to the dose ofhaggis, which with intravenous infusions of lipid that do not assess triglyceride concentrations from was comparable in all groups (mean 2 6 (2 5 to digestion and absorption.

fasting values. *p<005, http://www.bmj.com/ ***p

1634 BMJ VOLUME 297 24-31 DECEMBER 1988